Detection of breast cancer microcalcifications in digitized mammograms. Developing segmentation and classification techniques for the processing of MIAS database mammograms based on the Wavelet Decomposition Transform and Support Vector Machines.

摘要

Mammography is used to aid early detection and diagnosis systems. It takes an x-rayudimage of the breast and can provide a second opinion for radiologists. The earlieruddetection is made, the better treatment works. Digital mammograms are dealt with byudComputer Aided Diagnosis (CAD) systems that can detect and analyze abnormalities inuda mammogram. The purpose of this study is to investigate how to categories croppedudregions of interest (ROI) from digital mammogram images into two classes; normal andudabnormal regions (which contain microcalcifications).udThe work proposed in this thesis is divided into three stages to provide a conceptudsystem for classification between normal and abnormal cases. The first stage is theudSegmentation Process, which applies thresholding filters to separate the abnormaludobjects (foreground) from the breast tissue (background). Moreover, this study has beenudcarried out on mammogram images and mainly on cropped ROI images from differentudsizes that represent individual microcalcification and ROI that represent a cluster ofudmicrocalcifications. The second stage in this thesis is feature extraction. This stageudmakes use of the segmented ROI images to extract characteristic features that wouldudhelp in identifying regions of interest. The wavelet transform has been utilized for thisudprocess as it provides a variety of features that could be examined in future studies. Theudthird and final stage is classification, where machine learning is applied to be able touddistinguish between normal ROI images and ROI images that may containudmicrocalcifications. The result indicated was that by combining wavelet transform andudSVM we can distinguish between regions with normal breast tissue and regions thatudinclude microcalcifications.